In mps9506/dartx: Applies Drainage Area Ratio Method With Correction Factors

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  fig.retina = 2,
  out.width = "100%"
)

dartx

dartx is an R package to facilitate applying correction factors to streamflows calculated with the Drainage Area Ratio Method (DAR). dartx consists of a single function that calculates the streamflow probability (the non-exceedance probability), identifies the appropriate correction factor, and calculates the streamflow of an ungaged station with a user provided drainage area ratio.

Installation

You can install the the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("mps9506/dartx")

Currently there are no plans to submit the package to CRAN.

dartx requires the dplyr, fuzzyjoin packages. The examples below use ggplot2 and dataRetrieval packages.

Background

The DAR method is a simple method to estimate streamflows with the assumption that the relationship between streamflows at a gaged site and an ungaged site are equivelent to the ratio of the drainage areas. Asquith, Roussel, and Vrabel (2006) investigated the assumption of unity on the exponent parameter and multiplicative bias correction parameter in the DAR equation when applied to streams in Texas. The results includes suggested emprically derived parameters values for the DAR exponent at 54 percentiles of streamflow. When using these percentiles, bias correction approaches unity and can be assumed equal to 1.

More details are available in: Asquith, William H., Meghan C. Roussel, and Joseph Vrabel. 2006. “Statewide Analysis of the Drainage-Area Ratio Method for 34 Streamflow Percentile Ranges in Texas.” 2006–5286. U.S. Geological Survey Scientfic Investigations Report. U.S. Geological Survey. https://pubs.usgs.gov/sir/2006/5286/pdf/sir2006-5286.pdf.

Usage

At minimum the user must provide (1) a dataframe with streamflow from the desired reference gage and (2) the drainage area ratio as a number. dartx calculates the non-exceedance values, determines the correction value, and applies it to the DAR calculation. And example using the USGS dataRetrieval package is shown below.

## Load packages
library(dplyr)
library(dataRetrieval)
library(dartx)
library(ggplot2)

## Get drainage areas for two watersheds with dataRetrieval
siteInfo <- readNWISsite(c('08110100', '08109800'))
siteInfo$contrib_drain_area_va

## Download three years of streamflow data from 08110100
Davidson <- readNWISdv('08110100', "00060", "2015-01-01", "2017-12-31")
Davidson <- renameNWISColumns(Davidson)
head(Davidson)

The DAR is the area of the ungaged site divided by the gaged site.

dar <- siteInfo$contrib_drain_area_va[2]/siteInfo$contrib_drain_area_va[1]
dar

Apply the dartx function to get streamflows at the ungaged site.

output <- dartx(Davidson, flow = Flow, dar = dar)
head(output)

library(hrbrthemes)
output %>%
  rename(GagedFlow = Flow, EstimatedFlow = Q) %>%
  tidyr::gather(key = type, value = cfs, GagedFlow, EstimatedFlow) %>%
  filter(cfs > 0 ) %>%
  ggplot(aes(Q_percentile, cfs)) +
  geom_line(aes(color = type)) +
  scale_y_log10() +
  theme_ipsum_rc(axis = TRUE) +
  labs(title = "Flow Duration Curves", subtitle = "Measured and estimated flows at different sites",
       y = "Discharge (cfs)", x = "Streamflow percentile (Non-Exceedance Probability")

Make a quick comparison to the actual measured flows at site 08109800

Yegua <- readNWISdv('08109800', "00060", "2015-01-01", "2017-12-31")
Yegua <- renameNWISColumns(Yegua)

comparison <- Yegua %>%
  left_join(output, by = "Date") %>%
  filter(Q > 0)


ggplot(comparison) +
  geom_point(aes(log(Flow.x), log(Q)), alpha = 0.4) +
  geom_smooth(aes(log(Flow.x), log(Q)), alpha = 0.4, method = "lm") +
  geom_abline(intercept = 0, slope = 1) +
  coord_equal() +
  theme_ipsum_rc(axis = TRUE) +
  labs(subtitle = "Measured Flow vs Estimated Flow at the same site",
       x = "log Measured Flow",
       y = "log DAR Estimated Flow")